Agent skill security
Skills are instructions and supporting files that can change an agent's behavior after initial review.
Inspect the instruction surface
Review SKILL.md and supporting text for hidden Unicode, overriding instructions, encoded-payload indicators, referenced domains and command execution patterns. These are indicators that require human interpretation.
Keep private skills local
Use driftseal baseline on a local skill directory, retain the approved JSON and compare subsequent scans in CI. The scanner does not run skill instructions or follow instructions embedded in files.
What a clean result means
DriftSeal provides automated technical analysis and monitoring. It does not guarantee that software is secure. A clean DriftSeal result means only that no material issue or drift was detected by the active ruleset and coverage available during that scan.
Establish your baseline.
Inspect a public component or use the local CLI for private source.
Scan a component — free ↗